The GeForce RTX 5090 is the first thing that comes to mind when you think of the best graphics cards. However, you'd better pray the serial number on the inside of the metal bracket does not fade over time, since that could be a reason for Nvidia to reject warranty claims. At least one Redditor has shared a case of Nvidia reportedly denying a warranty claim because the bracket's serial number was unreadable.

The Redditor Willing-Avocado-4830, based in the UK, began experiencing black-screen crashes with a GeForce RTX 5090 Founders Edition graphics card under heavy load. The user sent the card to Nvidia for a warranty claim. After 30 days of radio silence, the owner finally received an email stating that their RMA request was rejected with no apparent explanation for the denial.

Seeking clarification, the user reached out to Nvidia for further information. According to the Redditor's recount, Nvidia support purportedly rejected the RMA because the serial number on the graphics card's metal bracket had faded. Like many gamers, Willing-Avocado-4830 claims never to have modified or tampered with the graphics card after installation. The fading serial number does not appear to be an isolated incident. Several other owners have reported similar issues in the Reddit discussion.

The metal bracket on graphics cards is susceptible to dust buildup, oxidation, and corrosion over time. This can cause it to look dirty, stained, or sometimes slightly discolored compared to when it was brand new. These same phenomena could have affected the serial number's legibility, especially since it is printed with ink onto the bracket's surface rather than etched for better durability. For comparison, the last-generation GeForce RTX 4090 Founders Edition had its serial number laser-etched onto the metal bracket.

One Redditor said that someone in the Nvidia Discord channel recommended cleaning the area with isopropyl alcohol to restore the serial number's visibility. However, that may not work if the ink is no longer present, as previous reports suggest the hot air expelled by the Founders Edition cooler could be erasing the serial number.

Surprisingly, Nvidia rejected a warranty claim solely because of a faded serial number on the graphics card’s bracket. The same serial number is clearly printed on a sticker on Nvidia's modular PCIe connector, which the chipmaker's technicians can verify when disassembling the GeForce RTX 5090. Moreover, you can pull the serial number from the graphics card through the command prompt with Nvidia SMI (System Management Interface), assuming the card is still working, which, in this case, it is. Willing-Avocado-4830 still had the original packaging and the original invoice from Nvidia U.K., so it would be easy to cross-reference the serial numbers.

Redditor hawok reportedly sent their GeForce RTX 5090 Founders Edition for RMA two months ago and encountered a similar situation. The serial number on the metal bracket was only partially visible. However, customer support seemingly accepted the RMA after the user sent photographs of the packaging that corroborated the serial number. It is unclear what Nvidia's policy is, or whether the chipmaker handles each case differently. We have reached out to Nvidia for clarification.

Some Redditors have suggested taking a photograph that shows the graphics card's serial number next to the packaging's serial number. Others jokingly recommend putting a piece of tape over the serial number. It might not be a bad idea, but you would need heat-resistant tape, something like Kapton tape.

GeForce RTX 5090 stock has practically disappeared from the U.S. market, with very few units starting at $6,500, 3.25X over the Blackwell flagship's MSRP. You cannot blame GeForce RTX 5090 owners for worrying. They already have a lot on their plates, dealing with the 16-pin power connector melting woes, and now also have to worry about fading serial numbers.

It's back: one of the best deals on a brand-new, current-generation, mid-range graphics card that packs 12GB of VRAM. There is a little hoop to jump through via Newegg to receive the discount, but it's more than worth it to slash $70 off the card and bring the price back down to its all-time low. AMD's Gigabyte Gaming Radeon RX 9070 GRE GPU is available at Newegg for $499. All you need to do is click on the "Extra Discount Available" link and enter your email address to receive the promotional code. Once you do that, you bring the price of this graphics card all the way down from its $569.99 list price. It's a very grim time for shopping for PC component upgrades, so to see an actual deal on a GPU that takes it below its MSRP in today's inflated market is a rare sight indeed.

● Check out this deal at Newegg

The RX 9070 GRE was initially made available to the Chinese GPU market and received a few changes from the original RX 9070 XT. The available VRAM was cut from 16GB to just 12GB of GDDR6, but it still sports bandwidth speeds of 18 Gbps on a 192-bit bus, producing 432 GB/s of memory bandwidth in gaming and applications. The Radeon RX 9070 GRE is still built on AMD’s RDNA 4 graphics architecture and uses the same Navi 48 GPU as the Radeon RX 9070 and RX 9070 XT. The RX 9070 GRE features a cut-down version of the Navi 48 chip with 48 compute units compared to the RX 9070's 56. Using just 220W total power draw, the RX 9070 GRE has an Identical power footprint to the standard Radeon RX 9070 GPU, making it quite power efficient and not needing a huge power supply to run.

We benchmarked the Radeon RX 9070 GRE in our extensive review, putting it through its paces in our suite of games. We found the card averages a cool 120 FPS at 1080p settings and 86.6 FPS at 1440p across our 11-game raster-only test suite. This is a good GPU choice for 1080p and 1440p gaming, with the RX 9070 GRE packing ample VRAM for high settings at 1080p, but pushing ultra settings at 1440p may use up the 12GB of VRAM. AMD's Radeon RX 9070 GRE sits just behind Nvidia's RTX 5070 in our results chart that you can view below.

If you're actively looking for a new GPU in the current PC component market, you're more than aware of the considerable price hikes across the GPU lineups from both AMD and Nvidia. To see a graphics card on sale for under its launch MSRP is a very rare sight, and I personally did not expect to see this kind of deal again outside of a large sale event like Prime Day or Black Friday. So if you're looking for a competent card for 1080p and 1440p gaming, then jump on the Gigabyte Gaming RX 9070 GRE for just $499.99.

Nvidia's fastest gaming graphics card, the RTX 5090, has been on a tear of price increases over the past several weeks. However, over the past week, the available inventory has dwindled. Now, you can only find the RTX 5090 from third-party sellers at online retailers like Newegg and Amazon, commanding anywhere from $6,500 to $9,500 (or even higher) for Team Green's best GPU.

At Newegg, the cheapest RTX 5090 is the MSI Ventus 3X that's available from Slava Computers (a relatively new seller with 239 ratings and a 2.8 out of 5 rating at the time of writing) for $6,449. On Amazon, you can get the Asus TUF Gaming OC for $6,395 from Joes Tech Shop, a seller with an 81% positive rating. However, among the most recent reviews are a string of one-star reviews about orders never being fulfilled. The cost goes much higher, as well. The first result for "RTX 5090" on Newegg, for example, surfaces the MSI Ventus 3X OC for $8,699.

In June, the median price for an RTX 5090 was $4,299. At the beginning of September, we logged the lowest online price at $5,199 in our GPU price tracker. Now, in less than two weeks, the available stock has completely disappeared online, and the options available from third-party sellers have ballooned in price once again.

Naturally, we don't recommend buying from one of these third-party sellers. The RTX 5090 is selling at a vastly inflated price, but more importantly, we've seen no shortage of scams around high-ticket items like the RTX 5090. In January, 42 Amazon customers were duped into a scam involving a $999 RTX 5090, instead receiving a fanny pack in place of the GPU. Earlier this month, an online seller scammed two buyers with the RTX 5090, selling phantom graphics cards with the core and memory removed on the secondhand market for prices near MSRP.

The one exception to RTX 5090 inventory is Micro Center, which still has cards available around the average selling price. At a local Micro Center we checked, the cheapest in-stock option was $4,299. Micro Center has exclusively sold graphics cards in-person for several years, insulating it from online buyouts like what appears to be going on right now.

As usual, the reason why there's so much demand for the RTX 5090 is fairly obvious: AI. The RTX 5090 holds the GB202 GPU and 32GB of GDDR7 memory. For context, Nvidia's RTX Pro 5000 48GB comes with the same GPU with a third fewer shader units enabled, and lower memory bandwidth compared to the RTX 5090, and sells for between $7,000 and $9,000. In that context, the RTX 5090 starts to look attractive for building an AI server, even at $5,000 (or more) apiece.

At the same time, the RTX 5090 remains the fastest gaming graphics card on the market according to the testing in our GPU benchmark hierarchy. That's even more true now with Nvidia's launch of DLSS 5 Neural Rendering, which the RTX 5090 can fully capitalize on (though, thankfully, we've found playable performance on lower-end cards in our extensive DLSS 5 testing).

We've reached out to Nvidia for comment on whether it has any plans to stabilize inventory. We'll update this story when we hear back.

The GeForce RTX 5090 is undeniably one of the best graphics cards money can buy. Banking on the fact that many already use it for AI, Nvidia has bolstered it with even more memory and launched it as the new RTX 5500 Pro Blackwell Workstation Edition. It offers comparable specifications to the GeForce RTX 5090 but distinguishes itself with 84GB of GDDR7 memory, 2.6X more than the Blackwell gaming flagship.

The RTX Pro 5500 features the GB202 silicon, the powerhouse die that powers other high-end mainstream and professional Blackwell graphics cards, including the GeForce RTX 5090, RTX Pro 6000, and RTX Pro 5000. In fact, the RTX Pro 5500 uses the same die as the GeForce RTX 5090, which means 170 Streaming Multiprocessors (SMs) are enabled out of a possible 192. As a result, the RTX Pro 5500 has 21,760 CUDA cores and offers performance comparable to the GeForce RTX 5090. The differentiator is the memory subsystem, where the RTX Pro 5500 excels.

Nvidia equipped the RTX Pro 5500 with 84GB of GDDR7 memory. This is the second time Nvidia has launched a graphics card with 84GB of memory, with the first being the China-exclusive RTX Pro 6000D. The RTX Pro 5500's memory capacity lets the Blackwell graphics card sit comfortably between the RTX Pro 5000 and the upgraded RTX Pro 6000, with 72GB and 96GB of GDDR7 memory, respectively.

Nvidia RTX Pro 5500 Workstation Edition Specifications

Nvidia achieved the 84GB memory capacity by outfitting the RTX Pro 5500 with 28 GDDR7 memory chips in a clamshell configuration. Each side of the PCB houses 14 24Gb (3GB) GDDR7 memory modules. They run at 25 Gb/s across a 448-bit memory interface, delivering up to 1,400 GB/s of bandwidth, which aligns with Nvidia’s official specification of 1,398 GB/s. In other words, the RTX Pro 5500 shares the same memory subsystem as the RTX Pro 6000D.

Samsung, Micron, and SK hynix only produce standard 28 Gb/s memory packages. Therefore, Nvidia downclocks the RTX Pro 5500's GDDR7 chips for its own reasons, whether to meet thermal and power targets or avoid cannibalizing higher-tier SKUs. From a memory-bandwidth perspective, the RTX Pro 5500 is not particularly impressive. Its bandwidth is only 4% ahead of the RTX Pro 5000, and it lags 22% behind both the flagship RTX Pro 6000 and the GeForce RTX 5090, which fully leverage 28 Gb/s GDDR7 memory chips.

To sum it all up, if RTX Pro 6000D and GeForce RTX 5090 had a child, it would be the RTX Pro 5500. Nvidia's main motivation was to launch the RTX Pro 5500 to fill the large performance and pricing gap between the RTX Pro 6000 and RTX Pro 5000. Also, the chipmaker makes substantially more profit by putting recycled GB202 into an RTX Pro 5500 than into a GeForce RTX 5090. By giving the former more memory with similar core specifications as the latter, Nvidia seemingly hopes to win more customers over to the RTX Pro 5500.

We reached out to both Nvidia and PNY for official pricing on the RTX Pro 5500, but neither company has responded. For reference, the flagship RTX Pro 6000 retails for $15,599, while an RTX Pro 5000 72GB starts at $9,209. Given the RTX Pro 5500's role as a bridge between the two models, it is reasonable to expect its launch price to fall somewhere within this range.

With the latest ROCm updates, AMD finally brought robust, official PyTorch and HIP SDK support to Windows for consumer GPUs, fully supporting the Radeon RX 7000 and the RX 9000 series. For native, supported frameworks, AMD on Windows is finally a viable reality, but what happens when you want to run a proprietary application, an older repository, or a specialized AI tool that absolutely refuses to support anything but NVIDIA's CUDA? That's where a new project, Speedstu's "CUDA-for-AMD-Windows," could save the day. It proves that running rigidly CUDA-exclusive workloads on AMD hardware in Windows is possible without virtualization or dual-booting.

To be clear, this project is not a brand-new runtime. Instead, it is a highly automated and reproducible PowerShell setup that bridges the gap between ZLUDA, the well-known, formerly AMD-funded translation layer, and AMD's native HIP/ROCm SDK for Windows. Through a series of clever scripts, the toolkit automatically detects the user's GPU architecture, grabs a specifically pinned version of ZLUDA (v6-preview.69), and, at least in theory, seamlessly maps it to the ROCm math libraries already present in Windows.

The result is that the developer successfully intercepted and mapped the CUDA driver API as well as the cuBLAS, cuSPARSE, and cuFFT libraries directly over to their AMD equivalents. As a proof-of-concept, the author even trained a 2.2-million-parameter PPO reinforcement-learning network end-to-end using unmodified CUDA libraries on an AMD Radeon RX 9060 XT, which happens to be the only officially supported GPU at this time.

Even with AMD's official ROCm support on Windows, the local developer community frequently runs into dependency hell when trying out experimental GitHub repos or specialized AI tools that hardcode CUDA as a requirement. For developers who want to experiment with these CUDA-only tools natively on their Windows daily driver machines without dealing with WSL2 passthrough issues or waiting for the original author to write a HIP port, this project offers a highly desirable translation pipeline. It acts as a sort of hacky adapter for software that stubbornly demands an NVIDIA card.

Benchmark testing included in the project's documentation offers some interesting findings. In a controlled A/B test running a 2.2M-parameter reinforcement learning workload on a Radeon RX 9060 XT, the "public upstream path," which relies purely on official ZLUDA releases and AMD's stock HIP SDK 6.4, achieved a median throughput of 13,278 steps per second (SPS). By contrast, an optional "recovered custom overlay" apparently built from salvaged legacy ZLUDA binaries ran slightly worse at 12,876 SPS, making it roughly 3% slower. While the clean official setup is faster, the author notes that "a later rewrite removed LibTorch/ZLUDA from PPO and achieved substantially higher throughput," indicating that there is still a performance hit for this stack of translators.

Now, before anyone declares the CUDA moat officially drained, it is crucial to set realistic expectations. First, this is a solo open-source project, not an enterprise-grade solution. The author is extremely transparent about its narrow scope, as crucial machine learning libraries like cuDNN, TensorRT, and NCCL do not resolve yet. This means compatibility is strictly workload-dependent; if your specific AI tool relies heavily on cuDNN, this setup will fail. Furthermore, it's worth pointing out that ZLUDA itself is currently being maintained as a "weekend hobby project" after losing its commercial backing a second time. Relying on this pipeline for production-level work remains a massive risk. This repository is a tinkerer's tool, not a corporate IT deployment strategy.

Despite these limitations, "CUDA-for-AMD-Windows" is pretty exciting, as it proves that the barrier to entry for running CUDA-exclusive software on AMD GPUs isn't an insurmountable hardware flaw, but a relatively tractable translation tooling problem. Because the project is entirely open-source, its potential extends far beyond this initial proof-of-concept. With community contributions, we could see expanded hardware detection and clever patches to get more stubborn CUDA libraries resolving properly.

If you believe that the 8-pin PCIe power connector is entirely immune to the problems that have plagued the newer 16-pin connector, think again. A recent incident suggests that even the humble 8-pin is not entirely immune to failure. According to a Radeon RX 7900 XTX owner on Reddit, the 8-pin connectors on a Thermal Grizzly WireView monitoring device suffered overheating damage while connected to the graphics card.

The user said that they experienced frequent freezes and black screens before discovering the damage. The affected connectors were the 8-pin connectors on the GPU side of the monitoring device, with visible signs of burning and melted plastic. The user also confirmed that the connectors were fully inserted, although the exact cause of the failure is still unknown.

WireView is essentially an adapter that sits between a graphics card and the power cables, allowing users to monitor various values like power consumption, voltages, current in amperes, minimum and maximum power consumption, and more. Rather than connecting the PSU cables directly to the GPU, the power passes through the monitoring device, thus adding another set of connectors and electrical contacts to the power delivery path. A poor electrical connection, increased contact resistance, a damaged connector, or an issue with the adapter itself could have caused the overheating damage.

Thermal Grizzly previously pointed to improper seating, backplate interference, or incorrectly aligned or soldered contacts as possible causes of connector damage, although the company is yet to identify the cause of this particular incident. The company has reportedly contacted the affected user and asked them to get in touch with its support team so the company can investigate the damaged WireView adapter. The investigation could potentially help determine whether the failure originated from the adapter, the connectors, the GPU, or a combination of factors.

While we are on the subject of 8-pin power connectors, just days ago we saw a heavily modified RTX 5090 successfully running on traditional 8-pin connectors. The modders replaced the card's original 16-pin connectors with three 8-pin connectors soldered directly to the PCB, using heavy-gauge cables and modified sense pins. The card reportedly pulled as much as 900W and reached 3,400 MHz, demonstrating how capable 8-pin connectors can be when properly implemented. However, this was an extreme hardware modification, not a stock configuration.

Discrete graphics card shipments for desktop PCs in Q2 2026 totaled 12.5 million units, the highest number since Q1 2022 despite record-high prices and cratering shipments of desktop CPUs, according to findings from Jon Peddie Research. The result highlights a broader trend that shows that unit sales of standalone GPUs for gaming have so far remained immune to rising prices, perhaps because gamers are expecting even higher prices in the coming quarters.

The industry shipped 12.5 million standalone graphics cards for desktop PCs in the second quarter of 2026, up around 5.9% sequentially and 7.8% year-over-year. 12.5 million add-in boards (AIBs) is the highest number of graphics cards sold in one quarter since the first quarter of 2022, when the industry shipped 13.38 million AIBs.

It is particularly noteworthy that 2026 is shaping up to be better for unit sales of desktop graphics boards than 2025 despite raising prices. For the first half of 2026, 24.3 million desktop AIBs were shipped, up significantly from 20.8 million graphics cards supplied in the first half of 2026. JPR analysts also note that only around 14 million desktop PCs were sold during the quarter, which — given an unusually high 89% attach rate — largely means that the majority of AIBs shipped during the quarter were aimed at gamers buying in retail and not at PC makers.

"Defying common wisdom, high-end AIB sales spiked as prices increased," said Jon Peddie, president of JPR. "Our theory is consumers rushed to buy AIBs before the prices went any higher, as the war in Iran is driving prices up in all segments."

Having shipped about 11.25 million discrete GPUs for desktop computers in Q2 2026, Nvidia remained the undisputed leader of the market with around 90% market share. AMD controlled roughly 8% of the market, shipping about one million discrete desktop GPUs, while Intel's share increased to 2% on shipments of several hundred thousand units. Meanwhile, Jon Peddie Research notes that market share changes were negligible during the quarter: AMD’s overall AIB market share decreased by -0.16% from the previous quarter, Intel's market share increased by 0.3%, and Nvidia's market share decreased by -0.1%.

For Nvidia, the quarter was particularly good as it sold the highest quantity of discrete GPUs for desktop PCs in a single quarter since Q3 2017, when it sold approximately 11.72 million units. By contrast, sales of AMD's standalone graphics cards have been floating below or around one million units per quarter for nearly four years now, with only three quarters being exceptions (Q3 2023, Q4 2023, Q4 2024). Still, one million is higher than the around 700 thousand discrete desktop GPUs the company sold in Q2 2025.

In total, Nvidia shipped approximately 17.365 million discrete graphics processors in the second quarter: roughly 11.25 million units went to desktops, and around 6.115 million units were installed into notebook and compact PCs. Since both AMD and Intel have quietly quit the market for standalone GPUs for mobile PCs, their shipments to this market segment were essentially zero.

Modders at the YouTube channel TecLab have created what is being called the world’s first RTX 5090 powered by traditional 8-pin PCIe connectors. Not only were the modders able to successfully power up the graphics card, but they also proved that the older connector is still quite capable of handling the extreme power demands of a modern high-end GPU.

The card chosen for the experiment was a Galax GeForce RTX 5090D HOF OC LAB XOC, which is a top-of-the-line model that features a beefy cooler as well as dual 16-pin power connectors. The modders essentially dismantled the entire graphics card along with the dual 16-pin power connectors, stripping it down to just the bare PCB. Next, three 8-pin power connectors were installed at the rear, with each connector soldered directly to the power regulators on the PCB using heavy-gauge cables. Additionally, the sense pins had to be rewired to ensure that the card could reach its maximum power load without any restrictions. To keep the temperatures in check, they used a liquid cooling system with a GPU block.

An 8-pin PCIe power connector is officially rated for 150W, although its actual maximum capacity is much higher. During the one-hour livestream, the modders tested the GPU by gradually increasing the power draw, starting at 400W and going all the way up to 900W using three 8-pin power cables. At its peak, the modified RTX 5090 pulled more than 120A, with clock speeds reaching as high as 3,400 MHz. Notably, the mod was limited to only the hardware level, with no software or BIOS tweaks performed.

To further test the capability of the 8-pin connectors, the modders tested the modified RTX 5090 with just two cables. The overall power draw dropped to 66A, while clock speeds remained at a respectable 3,200 MHz. Going a step further, the card was tested with just one cable, which surprisingly only dropped the figures down to 65A and 3,100 MHz.

Despite the PCI-SIG (PCI Special Interest Group) revising the 16-pin power connector (12V-2x6) to improve reliability, reports of melting GPU connectors continue to haunt gamers. The issue is particularly more common among high-performance models, though not limited to Nvidia RTX 50 and RTX 40 series, as certain AMD Radeon graphics with the 16-pin connector are also at risk. TecLab's experiment shows that GPU manufacturers may want to reconsider the tried-and-tested 8-pin connector and explore whether it could still be a viable option in high-end graphics cards.

A few years ago, Nvidia was expected to release the RTX 3070 Ti with as much as 16GB of VRAM instead of the 8GB we actually ended up getting. Only database listings ultimately indicated its existence until a physical prototype showed up, but we never really got to see it working. Fast forward to today, however, and Brazilian YouTuber fmklab, aka Fabian, has recreated the card on his own. Using an RTX 3070 PCB with a transplanted 3070 Ti core and GDDR6 chips from Samsung, fmklab broughall those parts together to achieve a 16GB memory pool.

The card was stripped down to its PCB, and the RTX 3070 GPU was removed and replaced with a 3070 Ti chip. The reason for the swap is that the RTX 3070 uses GDDR6 instead of the GDDR6X the retail version of the RTX 3070 Ti is equipped with, which presented an interesting challenge.

Fabian could've put all 16GB of GDDR6X memory on a 3070 Ti card, but it would've been far too expensive, so he swapped the Samsung-made HC14 (8Gb) GDDR6 chips that were already on the board with HC16 (16Gb) chips. That changed the VRAM capacity from 8GB to 16GB, but when the card booted, GPU-Z only recognized 8GB. The PCB's memory straps needed to be reconfigured in order for it to use the entire memory pool.

Using the Nvidia BIOS Reader tool, Fabian figured out exactly what straps to adjust and went ahead with the soldering process. Afterward, MATS properly recognized the full 16GB VRAM capacity with no errors. GPU-Z also showed the RTX 3070 Ti equipped with 16GB of GDDR6 memory, similar to the prototype that popped up years ago. But verifying that the card recognized the additional VRAM was just the first step.

To actually test his creation's performance, Fabian ran a local LLM on the GPU to see if it would allocate more than 8GB and found that it used nearly the entire 16GB pool. Call of Duty: Warzone ate more than 13GB of VRAM while paired with a Ryzen 7 9800X3D, exhibiting no signs of instability.

Fabian even linked a Unigine Superposition benchmark in the video's description as proof of the card's existence, where he scored 5,367 points, which is actually lower than most real 3070 Ti runs. That's probably because of the decreased memory bandwidth of GDDR6 versus that chip's stock GDDR6X, which is the biggest tradeoff for the higher capacity.

Overall, this was a successful experiment and possibly the only working 3070 Ti 16GB in the world, but there comes a point of diminishing returns. A 3070 Ti is still a powerful card, but it gets smoked by current-gen Blackwell options even with less memory. It would be much more reasonable to buy, say, an RTX 5060 Ti 16GB if you actually need better performance with some future proofing today, not to mention the superior efficiency, feature set, and after-sales support you'd get from a new Blackwell card.

There are now several mods enabling DLSS 5 on older Nvidia GPUs, along with even AMD's RDNA 4 cards. Similarly, Nvidia's multi-frame gen has also been ported to prior generations. It can get a bit confusing to keep track of all these developments, and that's where "DLSS Unlocked" comes in. It's a new mod from ShyVox that combines OptiScaler_DLSSNR with DLSS Enabler to bring DLSS 5 and DLSS MFG to all RTX GPUs: Turning, Ampere, and Ada.

DLSS Unlocked serves as a sort of all-in-one package for owners of RTX 20-, RTX 30-, and RTX 40-series looking to enjoy the latest Nvidia goodies on both Windows and Linux. It takes two separate mods developed by other community members and organizes them into one seamless, unified release. You download the executable from GitHub or Nexus Mods and just select the root directory of the game you want to try it on.

As for the requirements, the game must use the DirectX 12 API and already support both DLSS upscaling and frame generation natively. You need to provide your own patched DLSS 5 DLL as well, for legal reasons. The two mods comprising DLSS Unlocked also have their own peculiarities. OptiScaler_DLSSNR is simple; it just uses OptiScaler to add neural rendering on top, but DLSS Enabler works in a more complicated manner.

On RTX 20- and RTX 30-series GPUs, you'll be using AMD's FSR 3.1 to interpolate all of the frames instead of using Nvidia's pipeline. On RTX 40-series, the second frame is generated through DLSS-G (regular frame-gen), but the rest of the frames are handled by FSR 3.1 to get you up to 6X MFG. That handoff is itself possible thanks to another mod called DLSSG to FSR3. See how this web of tweaks can get a bit confusing?

You also need to keep in mind that this is a different approach compared to DLSS 5 Swapper, which enables DLSS 5 inside any game through ReShade (using RenoDX). For DLSS Unlocked, the game must already support DLSS 2/3 upscaling for it to be patched into DLSS 5 and unlock MFG. That's how OptiScaler_DLSSNR works, which DLSS Unlocked relies on.

None of these mods should be used with multiplayer games because the anti-cheat software can detect them as threats and potentially ban you. Ultimately, when you have DLSS Unlocked installed, a DLSS-compatible game should now have DLSS 5 neural rendering while having its frames interpolated up to 4x. You definitely need that, considering DLSS 5 comes with a massive compute cost that can purportedly even melt RTX 5090s.

Nvidia's GeForce RTX 50-series (codenamed Blackwell) graphics cards continue to rank among the best graphics cards available today. As a result, these highly sought-after Blackwell graphics cards have become prime targets for criminals seeking to scam unsuspecting buyers. Although scams involving hollow graphics cards are not new, their growing frequency and spread are cause for concern. The latest reports from Germany have revealed at least two buyers who fell prey to the same scam, reportedly from the same seller.

Hollow or phantom graphics cards may not be colloquial terms you hear every day. They basically refer to graphics cards that lack the silicon and memory chips. Criminals remove these components and sell non-functional graphics cards to their prey. Logically, you have no way of knowing these components are not present until you disassemble the graphics card, something that the average consumer does not do. Swindlers have run similar scams with fake memory modules without integrated circuits, or SSDs with microSD cards inside.

The first report, as recounted by respected hardware enthusiast Roman “der8auer” Hartung, uncovers a poor victim who bought a second-hand Palit GeForce RTX 5090 GameRock OC for around $2,090 (1,800 euros). The seller’s asking price was suspicious since a brand-new one starts at around $5,805 (4,999 euros) in Germany. Even for a used one, the dramatic price difference indicates a potential scam.

Many of these scams happen through online transactions. So the buyer went so far as to meet the seller in person, hoping a face-to-face interaction would ensure the graphics card was legit. It did not make a difference, though, because the buyer did not have a system to test the graphics card on the spot. In the end, it was a matter of taking the seller's word for it, and that turned out horribly.

A second report has emerged from a member of the Hardwareluxx forums, who reportedly fell victim to the same deceptive seller. Tragically, this individual ended up paying close to $2,613 (2,250 euros) for the phantom graphics card.

In both cases, the scammer put on a friendly facade and provided what seemed like legitimate paperwork to bolster the buyers' trust. For the first victim, the seller shared an invoice showing the purchase of two graphics cards, claiming to sell one because they no longer needed it. Similarly, in the second case, the scammer furnished an invoice to convince the buyer of the transaction’s legitimacy.

The use of realistic invoices is a growing trend among fraudsters. The first buyer uncovered that the serial number printed on the graphics card itself did not match the serial number listed on the packaging. The second buyer discovered that the serial numbers on the invoice and the packaging matched, but the seller deliberately scratched off the serial number on the graphics card.

The AI gold rush will not be over anytime soon, so we do not expect graphics card pricing to improve. This gives scammers more motivation to prey on buyers' desire for a graphics card and lure them in with prices that seem too good to be true. The two incidents serve as a cautionary tale for anyone trying to procure a graphics card, especially from second-hand sources.

Not that long ago, we reported that Nvidia was dusting off the blueprints for the RTX 3060, in its 12 GB form. At the time, we'd spotted it in stores for about $339.99, but just like with ever-climbing memory, hard drive, and SSD prices, two months passed is an eternity. The same cards are now selling for $489 new, and one particular specimen is the SRhonyra RTX 3060 12 GB Low Profile card, for $495.59.

• SRhonyra RTX 3060 12 GB Low Profile card - $495.59

This card and its price may raise more than a few eyebrows, but there are reasons why it exists. First off, it's a one-slot model, making it easy to put many of them to work in the same machine with relatively little concern for airflow. They have no power inputs and rely on 70 W delivered by the PCIe slot alone, eschewing the need for a high-end PSU and lots of cables. Third, the low-profile form factor makes it possible to place them in potent puny personal computers.

Attentive readers might surmise that one (or more) of these would be good candidates for an entry-level local LLM rig. That's precisely how SRhonyra is pitching the card, calling it "capable of local AI" and "running 7B [to] 13B LLMs." A standard-dimensioned, fully powered RTX 3060 is capable of drawing a maximum of 175 W, so it's fair to assume the performance of the diminutive variant will sit below that of its full-sized brethren, given it ought to only draw 70 W of juice from the PCIe slot.

Even then, it's likely that people interested in these cards are looking to use them either as secondary GPUs, or use more than one in the same box to be able to virtually pool their VRAM and use larger models than you'd otherwise be able to. The low power draw also means they're a simple, thoughtless drop-in to an existing system, whereas larger, more power-hungry cards require careful consideration with physical spacing (or lack thereof), power supply sizing, and ever-annoying cables.

Nvidia's DLSS 5 Neural Rendering technology has taken the PC gaming world by storm since its recent leak and then official rollout in recent days. In what might be one of the most impressive technical applications and mods of the feature yet, one developer has showcased DLSS 5 running on a second GPU to share the processing load, rendering the game on the first GPU before applying Neural rendering at the end of the frame, thus boosting the performance of neural-rendered frames.

Marcelo Guibout shared the demonstration online, with videos showing the process running on a cinematic video from The Blood of Dawnwalker, as well as Cyberpunk 2077. Guibout was quick to clarify that the videos are technical showcases, not benchmarks. However, they did share some performance figures. More exciting still, you can download the project from GitHub and try it for yourself. However, the technique does require a second display and doubles the display latency.

The above demonstration features a Ryzen 7 7800X3D, 32GB DDR5 setup with two Nvidia RTX 5060 Ti 16GB GPUs, with both cards using PCIe 5.0 x8, and a display attached to each card. The ReShade add-on, dubbed MGPU Bridge, reads each finished frame before applying Nvidia's Neural rendering. "Neural rendering happens at the end of the frame: it takes a finished frame and hands a finished frame back," they explain. "That is what makes it possible to pick it up and run it somewhere else. The add-on creates its own D3D12 device on your second GPU, sends each finished frame across to it, runs DLSS-NR there, and displays the result on that card's own monitor — so nothing has to come back. The render GPU does no neural work at all and runs cooler for the same reason."

In the TBOD demo at 1080p, they shared the following performance numbers:

Guibout clarified that DLSS super-resolution is running in every column of the test, with the rows showing each mode tested. The game rendered at the same internal resolution in all three columns, with the first column giving figures with DLSS 5's neural rendering switched off, representing the ceiling for performance. As you can see, running DLSS 5 on the second card in this dual-GPU setup drastically increases performance in every mode.

"The frame rates are not the finding. The slope is: what you gain by going from DLAA down to Ultra Performance, and how much of that available gain each arm keeps," Guibout explains. "Neural post-processing saturates whatever device it runs on, and it always runs at output resolution. Its cost barely falls as you drop the DLSS mode, while the render work collapses. On the render card it therefore eats a larger and larger share of every frame, and upscaling stops paying for itself: you keep about a third of what the machine actually had to give. Move it to the second card and you keep 86% of it."

Other benefits include temperature reductions on the rendering card, with the first GPU running 21 degrees cooler without the neural load.

Guibout is cautious to note this isn't a return to Nvidia's SLI technology, which would split the workload of frame rendering between two GPUs. They also noted plainly this is not Nvidia's vision for DLSS 5 and lacks the native integration of official DLSS 5 support.

You can see the original demo for TBOD below:

As commenters online have noted, the method seems more akin to using Nvidia's dedicated PhysX cards, with plenty of potential for gamers who have the capacity to run a dual GPU setup. The second demo features a Ryzen 5 5600 on a DDR4 system, so top-of-the-line hardware isn't required to make it happen.

The RTX 5090 is no stranger to controversy thanks to the incendiary nature of its 16-pin connector. When you add something as taxing as DLSS 5, which comes with a heavy increase in power draw, to an already dangerous mix, you're just cooking up a recipe for disaster. And it seems like this recipe has likely claimed its first victim. HardOCP forum member Erek claimed their RTX 5090 fell prey to the woes of the 16-pin connector yesterday, as it melted while testing DLSS 5 in NBA 2K27.

The irony is not lost on us that NBA 2K27 is the only game that officially supports DLSS 5 right now — everything else is modded — yet it still reportedly killed the user's MSI RTX 5090 Gaming Trio OC, which carries a 575W TDP. However, during testing, erek reported his GPU going past 600W, with GPU-Z showing it hovering above 610W consistently, not just as a spike. Before this, the card only consumed around 450W while gaming, implying that DLSS 5 added more than 150W to the power draw.

While the 16-pin connector is capable of carrying up to 600W of power, the PCIe x16 slot the card is plugged into also provides 75W on its own. So, the connector wasn't necessarily drawing over 600W alone. Tragedy struck right after as Erek's RTX 5090 started exhibiting signs of instability and eventually stopped working entirely. The owner smelled something burning and, sure enough, found the 16-pin connector pulling an age-old classic: melting.

The plastic on the connector had melted enough to bond with the cable, making it very difficult to separate the two. Once they did come undone, the damage was clearly visible, with at least five pins on the 16-pin cable scorched; one pin's housing had completely melted away. This was a yellow-tipped cable, too, meant to ensure it was plugged all the way in. Despite the fact that we only got a blurry picture of the connector on the GPU, the little yellow dot of melted plastic almost comically stands out amidst a sea of black and grey.

We can't confirm whether DLSS 5 was actually responsible for burning Erek's card. After all, an RTX 5090 meeting its maker is a monthly tradition at this point, and the cause is usually tied to a loose fit between the connector and the cable. However, our own testing does align with how the enthusiast came up with their DLSS 5 hypothesis. We saw the MSI Lightning Z variant of the 5090, which has 2x 16-pin connectors and a 1,000W XOC BIOS, pull 800W in Control with DLSS 5 enabled. Without DLSS 5, the same GPU pulled just 691W.

That wasn't even the most egregious difference; in Hogwarts Legacy, we went from 480W without DLSS 5 to a whopping 720W with it enabled. That's a 50% increase in power draw for just one feature. The same game made the Founder's Edition of the 5090 pull 417W without the neural rendering tech involved, and 547W with it toggled on. So even in the standard version of the GPU with a single 16-pin connector, there's a 31% jump in power consumption because of DLSS 5.

Considering Erek's 5090 model lands somewhere between those two, and given his own testing showing a severe power budget imbalance, it's not unreasonable to assume DLSS 5 was the culprit. Had the user's GPU been acting weird before they tried the new feature, we might be having a different conversation. But it's clear that Nvidia has a lot of work to do before DLSS 5 makes its way onto the RTX 40-series down the line. Performance improvements aside, the safety concerns stemming from poor efficiency also need to be addressed.

Nvidia launched the GeForce RTX 3060 (12GB) in early 2021 as an affordable mainstream gaming GPU with a 170W TDP, before re-releasing two months ago to offer relief during the ongoing component crisis. Throughout this time, not one person thought that the 3060 consumed too much power... except someone in China who decided to make a 75W version of the card. This blower-style variant has no 6- or 8-pin power connectors and runs entirely off the PCIe slot it'll be connected to, resulting in some expectedly underwhelming performance.

The RTX 3060 takes very well to undervolting and, therefore, can already be taken all the way down to just 100W at the cost of modestly reduced performance. A single PCIe x16 slot can provide up to 75W of power, so forcing a 3060 down to that number requires a shunt mod. This practice is usually associated with unlocking power limits on a GPU to chase overclocking feats where you decrease resistance, but in this case, you'd be increasing it.

This is a single-slot, low-profile card with a blower-style cooler about the size of a modern smartphone. You get just 1x HDMI and 1x DisplayPort in terms of connectivity. There is no fin stack present either, the PCB lacks a backplate, and the shroud is a thin metal sheet responsible for all the heat dissipation with a rudimentary heatsink in the middle.

A teardown of the GPU on BiliBili shows it's under-equipped from the inside, too. The memory chips have no thermal pads or paste on them. Instead, the metal shroud just touches the core directly to keep the entire thing cool. Remember that there's just one fan at the far end to blow hot air; there is no intake or proper airflow with positive pressure. So far, everything about this card, except its compact size, is looking subpar.

Once we get to testing, any skepticism is validated as the 75W RTX 3060 GPU is barely able to edge past 900 MHz, despite being rated for 1,770 MHz boost clocks. This results in a Time Spy score of just 4,821 points whereas a regular RTX 3060 easily scores upwards of 9,000 points in the same benchmark. Even older budget GPUs like AMD's iconic RX 580 and Nvidia's equally-popular GTX 1060 score more.

During the Time Spy run, the GPU hotspot also went past 90 degrees Celsius, proving that the cooler is barely performing if it's struggling to tame even a 75W card. Clearly, this GPU was suffering from severe thermal throttling on top of already having its power budget more than halved. One could make an argument that it's still an Ampere GPU, so it could make sense in low-power, compact systems. But there are now mini PCs with similarly performing or far more potent integrated graphics in 2026.

A Redditor has successfully connected a very powerful eGPU to the Steam Machine. Large_Customer_3840 briefly explains how they managed to add a potent AMD Radeon RX 9070 XT to Valve’s compact modern gaming desktop by fitting an M.2 to OCuLink adapter in the machine. That was cable-connected to an Aoostar EG01, which hosted the hulking triple-fan desktop graphics card. However, there remained some wrangling to do due to Steam Machine hardware limitations.

The idea outlined above sounds like a quick and simple hack to add a more powerful GPU to the Steam Machine, but there are a few limitations of the Valve hardware that need addressing. Firstly, this compact computer has just one M.2 PCIe 4.0 NVMe slot, which is used by the SSD that Valve supplies with the machine. Valve reckons folks can augment this storage using the microSD slot, just like Steam Deck gamers are accustomed to doing.

So, after fitting an OCuLink to M.2 adapter, Large_Customer_3840 had to find somewhere for the boot drive to live. The answer was simply to shove the boot drive into a USB-C to NVMe caddy and boot from that. Valve’s Steam Machine has a USB 3.2 Gen 2 (10 Gbps) port on the back which supports DP-Alt and Power Delivery. This is the fastest USB port it has, so it was used for the operating system.

A second snag with the M.2 to OCuLink adapter setup here is that there will need to be a case mod done so the cable can be threaded through the rear of the chassis. Large_Customer_3840 hasn’t made such a mod yet, and they are happy to have the system running open for now.

But many eGPU solutions like this, using the Aoostar EG01 in this instance, are inherently messy anyway. This older Aoostar eGPU dock has no enclosure, making it little more than a platform with a PCIe slot and supporting circuitry to host a desktop GPU, exposed, on your desk. Moreover, the EG01 doesn’t have a built-in PSU, unlike some models, so you also have a typically ugly PC power supply in view, too.

Looking beyond the aesthetic mess, the Redditor was pleased that “When I turned it on the 9070xt became the primary monitor and worked straight away. There was no hassle no selecting of graphic adapters.” The extra shot of performance delivered by one of the best graphics cards for PCs in 2026 was also welcomed. Apparently the OCuLink-connected Radeon RX 9070 XT eGPU ran Crimson Desert at over 100 frames per second using the High preset. It would have been interesting to know how the Steam Machine’s built-in discrete GPU handled the game at the same settings… As the Steam Machine basically runs a GPU similar to the Radeon RX 7600 – you can see our extensive Graphics Card Hierarchy data to see the expected difference between the RX 7600 and RX 9070 XT in various gaming scenarios. Online testing for Crimson Desert at the High preset yields frame rates of around 50 FPS, making the eGPU addition a sizeable performance buff.

Many AMD mini PCs come with USB4 ports nowadays, so it is a shame that the Steam Machine’s USB connectivity plateaus at USB 3.2 Gen 2 (10 Gbps) via a single USB-C on the rear. USB4, or better, is a more convenient hot-pluggable eGPU port opening up lots of other docking options for compact system users. It may offer worse bandwidth than OCuLink, but most folks find it an acceptable compromise. I’ve used both and was slightly annoyed by both, vowing not to be tempted into the eGPU world again until mainstream USB4 v2/Thunderbolt 5 solutions have arrived at affordable levels.

News of regional price increases for Nvidia graphics cards has been rolling in over the past little while, and those hikes have now arrived in the United States. We nearly spit out our coffee this morning while checking Newegg prices for Blackwell products. After months of painful but still relatively reasonable e-tail prices versus skyrocketing RAM and SSD costs, popular Blackwell GPUs are now eye-wateringly expensive. And those increases appear to be rolling out across other e-tailers, too.

As part of our ongoing research for the best graphics cards, we track the prices of every e-tail listing we can find for consumer graphics cards and calculate the median price of those products. We specifically track this figure as we feel it represents the price of a given graphics card model that you're most likely to find in stock, not stripped-down models that might be produced in limited volume to hit an artificially low MSRP.

With that, the theoretically entry-level RTX 5060 is now $469.99 at the midpoint of current prices, which is now two rungs up the MSRP ladder compared to its $299.99 launch MSRP. Just a couple of months ago, 5060s were selling for a median $369.99, or just below the RTX 5060 Ti 8GB’s $379.99 launch MSRP. Now, the cheapest GDDR7 Blackwell card costs more than the RTX 5060 Ti 16GB’s $429.99 launch price.

RTX 5060 Ti 8GB cards used to be among the least marked-up Blackwell parts thanks to the fact that their performance and VRAM capacity was out of line with their high $379.99 launch MSRP, but they’re now headed up the escalator like their stablemates. The median 5060 Ti 8GB now costs $529.99, which is about 13-15% more expensive than a couple of months ago.

The RTX 5060 Ti 16GB now commands an astounding $799.99 median price, a jaw-dropping 38% more expensive than the $579.99 midpoint we last calculated. The 5060 Ti 16GB was already far too expensive to recommend for gaming at that price, and the new markup suggests that it’ll only be of interest to local AI explorers trying to get the most VRAM they can on a Blackwell card for under $1000. Pour one out for what used to be the best entry-level enthusiast GPU we recommended.

The RTX 5070 was another one of the last gaming holdouts near its MSRP thanks to strong competition from the RX 9070 16GB, but the midpoint of prices for 5070s has now leaped an incredible 29% over our last survey, to about $850-$900. Even comparing lows to lows, prices for the cheapest 5070s have jumped about 20%. That takes this card entirely out of the midrange running and positions it closer to the much faster RTX 5070 Ti, which also has 16GB of VRAM to play with.

The hikes appear to have hit the middle of the Blackwell lineup the hardest, as the midpoint of RTX 5070 Ti prices is the same as it was during our last check-in. RTX 5080s haven’t gotten substantially more expensive than they have been, either, as prices for those cards have always been highly elevated compared to their $999 MSRP. And the RTX 5050’s price has barely moved today, either, hovering near the $300 it’s maintained since around the beginning of the year.

These increases haven’t been matched by hikes on the AMD side—yet. Heavy emphasis on yet. We’re only seeing single-digit percentage increases in RDNA 4 card prices compared to our last survey, although Radeon RX 9000-series cards appear to be affected by the silicon supply crunch in other ways.

The assortment of available RX 9070 16GB cards is perhaps a bit smaller than it’s been in the past, while the cut-down RX 9070 GRE is available in abundance around its $549 MSRP, suggesting that card has taken over the true midrange role the plain 9070 could never quite manage at the same MSRP.

We felt that the GRE’s price was high at launch, but AMD likely has a better crystal ball for silicon supply chain trends than we do, as the GRE now offers incredible bang for the buck compared to the RTX 5070’s new sticker.

The RX 9060 XT 8GB isn’t completely dead yet, but only one XFX 8GB model remains readily available at e-tail for $399. The RX 9060 XT 16GB’s median price has slightly risen to $474.99, and the RX 9070 XT now sits 5% higher than our last check-in at a median of $799.99.

All told, these Blackwell price hikes are another body blow for a DIY PC component market that’s already reeling from sky-high RAM and NAND prices. Graphics cards had until recently been one of the less hiked-up component categories in a DIY PC’s bill of materials compared to the pre-AI times, but that period of relative solace is well and truly over if you want access to Nvidia’s hardware and software stack.

If you’re a PC gamer, there’s no good news here. We’ll have to see whether there’s a similar price spike waiting in the wings for Radeon cards in the coming days, or whether this is the sad, sorry new normal.

Nvidia is reportedly testing variations of its upcoming Rubin Ultra accelerator with less memory due to concerns it won't be able to source enough HBM. Some versions include just 192 GB of memory and use HBM4 instead of HBM4E, as originally announced, according to The Information. The report confirms an earlier comment from firm SemiAnalysis about a potential Rubin Ultra memory downgrade.

We first saw Rubin Ultra in the flesh earlier this year at GTC, where Nvidia showed off a compute tray housing four compute chiplets alongside 1 TB of HBM4E memory. The accelerator is part of Nvidia's Kyber NVL144 design, which is set to roll out in 2027. SemiAnaylsis reported that the rack was delayed to 2028. "Our roadmap is intact," said Nvidia to Tom's Hardware in response, though the company made no clarification on if the delay was real or not. We've reached out to Nvidia regarding this latest report.

According to The Information, Nvidia is testing versions of Rubin Ultra with 192 GB or 256 GB of memory, as well as versions that use fewer than the 16 announced memory stacks. Perhaps most importantly, Nvidia is reportedly testing with HBM4, not HBM4E as originally announced. Along with the traditional improvements we see in each new HBM generation, HBM4E is unique in that it offers a customizable base logic die. Last year, Micron announced a partnership with TSMC to manufacture the base die and allow customers to tweak the logic die based on their needs.

The complexity of HBM4E has reportedly caused a strain on supply, with memory manufacturers unable to keep pace with Rubin Ultra's rollout. At least three lower-memory designs have been tested by Nvidia, according to the report, though we don't have a full picture of details on those prototypes. The report claims testing with HBM4, as well as 192 GB and 256 GB configurations, though it makes no mention of the number of compute dies, nor the memory type for each tested capacity.

The number of dies is important. In June, reports circulated that Nvidia cancelled its quad-die Rubin Ultra design due to manufacturing complexities. Although Nvidia has yet to comment, reports at the time suggested Nvidia would move ahead with a dual-GPU Rubin Ultra. In such a case, less memory would make more sense. Even with a dual-die Rubin Ultra, the quoted capacities are lower than expected. Each base Rubin GPU currently ships with 288 GB of HBM4.

It's clear Nvidia is trying to get ahead with memory in a world where agreements have been signed multiple years into the future. Nvidia has several of its own agreements. In June, the company announced a partnership with SK hynix to develop next-generation memory technology, which includes HBM, but also LPDDR5X and DDR5. In July, Nvidia expanded that partnership with a $500 billion strategic relationship that includes a long-term memory supply agreement with SK.

Although Nvidia is considering lower-memory configurations, one Nvidia customer told The Information that per-GPU memory isn't a top concern, valuing the relationship with Nvidia over the long term.

Memory shortages are touching nearly every design currently on the market, though enterprise systems packing HBM are particularly vulnerable. Last week, Digitimes reported that Samsung, SK hynix, and Micron have sold through their HBM capacity through 2027. Last month, SK Hynix CEO Kwak Noh-jung said 2027 will be the "worst year" for the memory shortage, with supply constraints lasting through 2030.